1

Vienna LTE Simulators LTE-A Link Level Simulator Documentation, v1.1

Institute of Telecommunications Vienna University of Technology, Austria Gusshausstrasse 25/389, A-1040 Vienna, Austria Email: {sschwarz, jcolom, msimko}@nt.tuwien.ac.at Web: http://www.nt.tuwien.ac.at/ltesimulator

Abstract This document contains documentation on how to use the Long Term Evolution Advanced (LTE-A) Link Level simulator [1] from the Vienna LTE simulator suite, as well as some insight on its structure and the assumptions that were made while developing it. This simulator is an extension of the previously published Vienna LTE Link Level Simulator. This document gives an introduction on how to use the simulator. The concept and the structure of the simulator is described in more detail in [2].

I. F OREWORD The LTE-A link level simulator is published under a non-commercial academic use license. Please make sure that you understand the terms and conditions of the license before you use any of the available software packages. Would you require a license different to a non-commercial academic one please contact Stefan Schwarz, Michal Simko or Josep Colom Ikuno. The detailed license agreement for the LTE-A Link Level simulator can be found in Section XVIII. Please read the license agreement carefully, as parts of the code are under the GNU Lesser General Public License [3], and the MIT License [4]. II. RUNNING THE SIMULATOR FOR THE FIRST TIME The LTE-A Link Level simulator is provided with several test simulation scenarios that can be used to verify that the simulator runs in the expected way. In the ./examples directory you can execute the LTE_sim_batch_quick_test M ATLAB script. This script provides a means to test the performance of a single user Long Term Evolution (LTE) transmission on an uncorrelated TU channel for several transmission modes [5] and antenna congurations (number of transmit antennas Nt times number of receive antennas Nr ). All simulated transmission modes (Single-Input Single-Output (SISO), Transmit Diversity (TxD), Open Loop Spatial Multiplexing (OLSM) and Closed Loop Spatial Multiplexing (CLSM)) utilize User Equipment (UE) feedback to adapt important transmission parameters (code rate, modulation alphabet, Multiple-Input Multiple-Output (MIMO) preprocessing) to the channel quality (see [6] for details). The script produces plots of the cell throughput obtained with the different transmission modes. To verify their correctness, these plots can be compared to the gures provided in the subdirectory ./LTE quick test results of the directory ./examples or to the curves shown in Figure 1. The result of the simulation is a group of result les that is stored in the ./results folder. Details to the simulation setup can be found in [7].
10 9 8 Throughput [Mbit/s] 7 6 5 4 3 2 1 0 10 5 0 5 10 15 SNR [dB] 20 25 30 35 SISO Transmit diversity OLSM CLSM

Fig. 1.

Plots resulting from running the LTE_sim_batch_quick_test.m M ATLAB script

Parameter Number of UEs Bandwidth HARQ Retransmissions Channel type Filtering Receiver type Simulation length Transmit modes

Value 1 1.4 MHz 0 TU uncorrelated Block Fading Zero Forcing 1000 subframes SISO, TxD (Nt Nr = 2 2), OLSM (2 2) and CLSM (4 2)

TABLE I BASIC SETTINGS USED FOR THE LTE_ S I M _ B A T C H _ Q U I C K _ T E S T . M M ATLAB SCRIPT

Parameter Number of UEs Bandwidth HARQ Retransmissions Channel type Filtering Receiver type Simulation length Transmit modes Precoder feedback

SUMIMO settings Value 1 1.4 MHz 0 TU uncorrelated Block Fading Zero Forcing 1000 subframes CLSM 8 1, 8 4 and 8 8 Standard dened codebook

MUMIMO settings Value 5 1.4 MHz 0 at Rayleigh uncorrelated Block Fading Zero Forcing 1000 subframes Non-standard dened precoding (ZF beamforming) 4 1 Perfect channel knowledge and RCVQ

TABLE II BASIC SETTINGS USED FOR THE LTEA_ S I M _ B A T C H _ Q U I C K _ T E S T _SUMIMO. M AND LTEA_ S I M _ B A T C H _ Q U I C K _ T E S T _MUMIMO. M M ATLAB SCRIPTS

The most important settings used in the LTE_sim_batch_quick_test M ATLAB script can be found in Table I. To test the LTE-A capabilities of the simulator, the two M ATLAB scripts LTEA_sim_batch_quick_test_SUMIMO and LTEA_sim_batch_quick_test_MUMIMO are provided. The SUMIMO test script compares the single user throughput performance of different LTE-A antenna congurations (eight transmit antennas), when utilizing the standard dened precoder codebook [8]. Again the transmission parameters are adapted in accordence with the channel quality experienced and fed back by the UE (assuming instantaneous feedback), see [9] for details. In Figure 2, the SUMIMO throughput for three different antenna congurations is shown. Running the script le LTEA_sim_batch_quick_test_SUMIMO should produce similar results. The simulation settings utilized in these simulations are given in the second column of table Table II.
35 8x1 8x4 8x8

30

25 Throughput [Mbit/s]

20

15

10

0 10

10 15 SNR [dB]

20

25

30

35

Fig. 2.

Plots resulting from running the LTEA_sim_batch_quick_test_SUMIMO.m M ATLAB script

In addition to standard dened precoding, LTE-A also supports non-standard dened precoding. As an example, Zero Forcing (ZF) beamforming for single receive antenna UEs is implemented in the LTE-A simulator. This precoding method spatially orthogonalizes the transmission to different UEs, by appropriately forming antenna beams. The precoders are either computed from perfect channel knowledge, or utilizing Channel Vector Quantization (CVQ), from quantized (imperfect) channel knowledge. For details on the utilized feedback and precoding methods, see [10]. Figure 3 shows the obtained cell throughput versus SNR for the two considered feedback methods, assuming a frequency at Rayleigh fading channel

model. Note that no precoder interpolation is currently implemented. Therefore the precoding method will not perform well in frequency selective channels, because only one channel value per Resource Block (RB) is fed back. Running the script LTEA_sim_batch_quick_test_MUMIMO in the ./examples directory allows to reproduce these results. The most important simulator settings for these simulations are summarized in the last column of table Table II.
20 18 16 Throughput [Mbit/s] 14 12 10 8 6 4 2 0 0 5 10 15 SNR [dB] 20 25 30 Perfect channel knowledge RCVQ based on LTEA codebook

Fig. 3.

Plots resulting from running the LTEA_sim_batch_quick_test_MUMIMO.m M ATLAB script

III. MEX FILES Computation-intensive parts of the LTE-A link level simulator are implemented in C and used in the simulator by means of MEX les. The source code for the given les can be found in the /C-source folder under the simulator root folder. Please not that some of the functions there are licensed under other license terms. Please check Section XVIII for more details. The MEX les distributed with the simulator release are the Windows 64-bit version. Should you require them for any other platform you recompile them by using the LTE_aux_mex_files script. You can nd more information on how to use/write/compile MEX les here. IV. RUNNING YOUR OWN SIMULATIONS While the LTE_sim_batch_quick_test, LTEA_sim_batch_quick_test_SUMIMO and LTE_sim_batch_quick_test_MUMIMO M ATLAB scripts are adequate to get familiar with the basic structure of the simulator, you may need more exibility than the one offered with the quick test scripts. For this purpose, you can check the LTE_sim_batch.m script. It provides the means to set basic simulation parameters (SNR range, number of subframes simulated, choice of simulation conguration) and calls all necessary scripts in the appropriate order. Below you can nd a list of exemplary parameters that you may want to congure in the batch le: cqi_i: set of Modulation and Coding Schemes (MCSs) that are used for the simulation. In [11], 15 different Channel Quality Indicators (CQIs) are specied. If you want to simulate with all possible CQIs, just set the cqi_i loop to run over [1:15]. Note: depending on the settings you choose in LTE_load_parameters, the appropriate CQI value for the current channel conditions may be automatically determined and set during simulation. N_subframes: the length of the simulation, or how many subframes (Transmission Time Intervals (TTIs)) are simulated for each value of cqi_i and SNR_vec. SNR_vec: a vector containing the average Signal to Noise Ratios (SNRs) that will be used for each simulation run. All users will experience the same average SNR. See Section IX for a more detailed denition of the SNR. Alternatively, for simulations with multiple UEs, you can set SNR_vec as a matrix, where each row then corresponds to the SNR values of a single user. Note: the SNR matrix is then simulated column-wise, meaning that in each simulation run, a column denes the current average SNRs of the different users. LTE_load_parameters: load the parameter le that congures the simulator. A more detailed description of the available conguration parameters can be found in Section V. Six basic precongured options are given that can be directly used by appropriately setting the value of the variable Simulation_type, thus skipping the conguration of the rest of the parameters. 1) SUSISO: Single-user SISO simulation 2) MUSISO: SISO simulation with multiple users (precongured to two users)

3) SUMIMO: Single-user MIMO simulation utilizing CLSM 4) MUMIMO: Multi-user (four users) MIMO simulation. In this simulation, the columns of the standard dened precoding matrices are shared by different users in a round robin fashion (no sophisticated scheduler is currently implemented). The number of transmit and receive antennas equals 4. 5) LTE-A MUMIMO: Non-standard dened precoding (ZF beamforming) based MUMIMO with 5 single antenna users and four transmit antennas. 6) LTE-A SUMIMO: SUMIMO based on the LTE-A specied single user codebook (enables up to eight layers for a single user). The number of transmit antennas is precongured to 8, the number of receive antennas to 4. LTE_sim_main: main routine of the simulator. Generate the output lename and save the results in a .mat le. V. S IMULATION PARAMETERS

Below you can nd a list of the parameters that can be congured in the LTE_load_parameters le: A. General parameters

LTE_params.nUE: number of UEs to simulate. LTE_params.nBS: number of eNodeBs (cells) that will be simulated. LTE_params.uplink_delay: the delay the uplink experiences. It applies to ACKnowledgments (ACKs), CQI, Precoding Matrix Indicator (PMI) and Rank Indicator (RI) reports. An integer number 0. Useful if you want to experiment with scheduling algorithms or feedback strategies. LTE_params.show_plots: whether plots are shown during the simulation or not. LTE_params.trace_subcarrier_SNR: if set to true, a trace of the subcarrier SNRs is generated and stored in the simulation_results object. LTE_params.N_seed_reset: resets the random number generator seeds to a new value after LTE_params.N_seed_reset subframes. This is used for the case where a time-correlated channel is generated (modied Rosa Zheng model [12], [13]). When low speed channels are evaluated, a too-long simulation would be needed in order to obtain statistically meaningful results. Thus, to avoid such situations, the channel can be reset to a new seed every N subframes. LTE_params.carrier_freq: carrier center frequency [Hz] LTE_params.Bandwidth: system bandwidth. Allowed values are 1.4 MHz, 3 MHz, 5 MHz, 10 MHz, 15 MHz, and 20 MHz. This bandwidths are equivalent to 6, 15, 25, 50, 75, and 100 RBs respectively. Carrier aggregation (to enable bandwidths 20 MHz) is currently not supported by the LTE-A simulator. LTE_params.HARQ_processes: number of parallel Hybrid-ARQ (HARQ) processes. The maximum value, according to [14] is 8. LTE_params.max_HARQ_retransmissions: maximum number of HARQ retransmissions, not including the original transmission. Valid values are 0, 1, 2 or 3. Higher numbers will give you an error, as the rate matching is not dened for retransmission numbers (rv_idx) higher than 3. LTE_params.SubcarrierSpacing: in Hz, 15 kHz, a 7.5 kHz subcarrier spacing is also possible (just for MBSFNbased multicast/broadcast transmissions). Tests were so far performed using a 15 kHz spacing, so the 7.5 kHz spacing is not thoroughly tested. LTE_params.CyclicPrefix: cyclic prex length [15]. Either normal or extended for MBSFN-based multicast/broadcast transmissions. LTE_params.simulation_type: the simulator is capable of using the M ATLAB Parallel Toolbox in order to speed up simulations by using parfor loops. If you happen to have the Distributed Computing Toolbox, you will also be able to make use of it by using this option. Set this variable to parallel or normal to parallelize the SNR loop in LTE_sim_main or just perform a single-core simulation. Keep in mind that some modications you do to the code may not work in the parallel version or may directly cause it not to run. LTE_params.simulate_with_all_zero_sequences: true if you want that the transmitted data is an all-zero sequence (useful for interleaver testing and debugging). LTE_params.introduce_frequency_offset: whether you want a frequency offset to be introduced. Additonal carrier offset options are also present in the UE parameters conguration Section V-C. LTE_params.random_noise_seeding: whether the seed for the random number generator that generates the noise is set (allows for repeatability of the noise realizations). LTE_params.noise_seed: Only used if the upper variable is set to true. Integer number that sets the random number seed of the noise random number generator. LTE_params.usePBCH: whether space for the physical broadcast channel is reserved inside the resource grid (no data is transmitted on these resource elements). LTE_params.usePDCCH: whether space for the physical downlink control channel is reserved inside the resource grid.

LTE_params.trafficmodel.usetraffic_model: whether users generate trafc according to prespecied trafc models (RAN R1-070674) or a full buffer situation is assumed. Currently only a round robin scheduler (RoundRobin Trafc) is supported in combination with the trafc models.

B. Channel matrix source

LTE_params.channel_matrix_source: Controls the generation of the channel matrix trace. generated to generate it every time. trace to load it from a trace. LTE_params.store_channel_trace: Set to true or false. If mode is generated, the channel trace will be saved at the end of the simulation. LTE_params.channel_matrix_tracefile: lename of the trace le where the generated channel matrix trace is stored. Only applicable if trace mode is used (if the mode is set to trace, the channel matrix is already read from a trace, so it is meaningless to save it again in another trace).

C. UE, eNodeB and UE feedback specic parameters

LTE_params.UE_config.LLR_clipping: species the LLR clipping level. LTE_params.UE_config.turbo_iterations: Number of iterations of the turbo decoder. Set by default to 8. LTE_params.UE_config.N_soft: Denes the total number of soft channel bits available for HARQ processing (TS 36.306 4.2.1.3 [16]). LTE_params.UE_config.channel_estimation_method: Currently the following channel estimators are available: PERFECT, LS, or MMSE for LTE. LTE-A (transmission modes 8 and 9) only supports perfect channel knowledge at the moment. LTE_params.UE_config.channel_interpolation_method: linear, cubic, spline, sinc_freq, sinc_time, or T-F. For fast fading, linear, cubic, and v4 are available. LTE_params.UE_config.autocorrelation_matrix_type: type of autocorrelation matrix. Either ideal or estimated. LTE_params.UE_config.user_speed: channel speed (m/s). LTE_params.UE_config.realization_num: number of channel realizations. Used for averaging to obtain the channel autocorrelation matrix. LTE_params.UE_config.realization_num_total: rst xy number of channel realizations are used just for the estimation of the autocorrelation matrix. LTE_params.UE_config.CDD: Cyclic Delay diversity 0: zero delay CDD (3GPP TS 36.211-820 Section 6.3.4.2.1 [15], page 37) 1: small delay CDD (3GPP TS 36.211-820 Section 6.3.4.2.1 [15], page 37). Although in the newest standard version this is not dened anymore 2: large delay CDD (3GPP TS 36.211-820 Section 6.3.4.2.2 [15], page 38) LTE_params.UE_config.mode: the transmission modes are dened in TS 36.213-a01 Section 7.1 [17]. Following transmission modes are implemented (for details see Section VIII): 1: single antenna. 2: TxD . 3: OLSM. 4: CLSM. 5: Multiuser MIMO (only a pseudo implementation is available, no meaningful scheduling). 8: LTE-A non-standard dened precoding (ZF beamforming implemented as an example) 9: LTE-A SUMIMO (up to eight layer transmission). LTE_params.UE_config.nRX: number of receive antennas at the UE. LTE_params.UE_config.carrier_freq_offset: carrier frequency offset normalized to subcarrier spacing (not yet implemented). LTE_params.UE_config.perfect_freq_sync: whether the UE is perfectly synchronized in frequency. LTE_params.UE_config.rfo_correct_method: receiver frequency offset correction method. Either none, subframe, or FIR. Not yet implemented. LTE_params.UE_config.receiver: either SSD for a Soft Sphere Decoder, ZF for a ZF receiver and MMSE for a Minimum Mean Square Error (MMSE) receiver. LTE_params.UE_config.PMI_fb_granularity: granularity of PMI feedback in multiples of resource blocks LTE_params.UE_config.CQI_fb_granularity: granularity of CQI feedback in multiples of resource blocks LTE_params.UE_config.PMI_fb: whether PMI feedback is activated or not (true/false) LTE_params.UE_config.RIandPMI_fb: whether RI feedback is activated in addition to PMI feedback (true/false)

LTE_params.UE_config.CQI_fb: whether CQI feedback is activated LTE_params.UE_config.predict: whether channel prediction is activated for the feedback calculation; the channel prediction algorithm is simple linear extrapolation on every subcarrier; LTE_params.UE_config.SINR_averaging.averager: denes the Effective Signal to Interference and Noise Ratio Mapping (ESM) averager used. Possible values are EESM and MIESM. LTE_params.UE_config.SINR_averaging.EESMbetas: denes the calibration factors required for Exponential Effective Signal to Interference and Noise Ratio Mapping (EESM). Those values are obtained from extensive training simulations and should not be changed. LTE_params.UE_config.SINR_averaging.MIESMbetas: denes the calibration factors required for Mutual Information Effective Signal to Interference and Noise Ratio Mapping (MIESM). Those values are obtained from extensive training simulations and should not be changed. LTE_params.UE_config.SINR_averaging.MCSs: denes the used MCS set. Modication of these values might lead to unexpected behaviour, as the calibration factors (EESMbetas and MIESMbetas) are not trained for other MCSs. LTE_params.BS_config.nTx: number of antennas at the eNodeB. LTE_params.feedback.ignore_channel_estimation: whether the channel estimation mean square error is taken into account during the feedback calculation or not. If channel estimation is activated, the option should be set to true, otherwise the performance of the system migth be overestimated. LTE_params.feedback.channel_averaging: whether channel averaging is used during feedback calculation or not. If set to true just a single average channel value per resource block is used to compute the feedback for complexity reduction. Especially in 4 4 systems this degrades the performance of the feedback method. LTE_params.feedback.codebook: channel vector quantization codebook used for ZF beamforming (transmission mode 8). PERFECT provides perfect channel knowledge to the eNodeB (subsampled to one value per RB), LTE utilizes the LTE precoder codebook to quantize the channel. LTE_params.feedback.codebook_size: size of the codebook (in number of code vectors) if a DFT or RANDOM codebook is employed. LTE_params.feedback.CQI_quantization: whether quantized or unquantized channel quality (receive SNR) feedback is provided.

D. Channel model parameters Parameters that congure how the channel is generated and the signal ltered. LTE_params.ChanMod_config.filtering: BlockFading (channel is constant during one subframe) or FastFading (Fast Fading is currently not supported for LTE-A TX modes 8 and 9). LTE_params.ChanMod_config.interpolation_method: the channel interpolation method for the channels which are generated in the simulator. Either shift_to_nearest_neighbor for nearest neighbor interpolation or sinc_interpolation for sinc interpolation, which is more precise. Necessary if the channel sampling rate is not equal to the sampling rate of the transmit signal. LTE_params.ChanMod_config.sin_num: species the number of sin realizations used for the modied rosa-zheng model [12], [13]. LTE_params.ChanMod_config.type: species the type of channel used. The available ones are: AWGN: Additive White Gaussian Noise channel. flat Rayleigh: temporally uncorrelated frequency at Rayleigh fading channel. flat Rayleigh corr: temporally correlated frequency at Rayleigh fading channel. Tap-delay based models: PedA, PedB, PedBcorr, VehA, VehB, TU, RA, and HT [18], [19]. Externally-generated channel coefcients: winner_II. Uses the publicly-availabe Winner II implementation to generate the channel coefcients [20]. The following parameters can be congured when using the Winner II channel model. LTE_params.ChanMod_config.winner_settings.Scenario: 1=A1, 2=A2, 3=B1, 4=B2, 5=B3, 6=B4, 7=B5a, 8=B5c, 9=B5f, 10=C1, 11=C2, 12=C3, 13=C4, 14=D1 and 15=D2a. LTE_params.ChanMod_config.winner_settings.PropagCondition: LOS or NLOS. LTE_params.ChanMod_config.winner_settings.SampleDensity: number of time samples per half wavelength LTE_params.ChanMod_config.winner_settings.UniformTimeSampling: use same time sampling grid for all links (yes or no). LTE_params.ChanMod_config.winner_settings.FixedPdpUsed: nonrandom path delays and powers (yes or no). LTE_params.ChanMod_config.winner_settings.FixedAnglesUsed: nonrandom AoD/AoAs (yes or no).

LTE_params.ChanMod_config.winner_settings.PolarisedArrays: usage of dual polarised arrays (yes or no). LTE_params.ChanMod_config.winner_settings.TimeEvolution: usage of time evolution (yes or no). LTE_params.ChanMod_config.winner_settings.PathLossModelUsed: usage of path loss model (yes or no). LTE_params.ChanMod_config.winner_settings.ShadowingModelUsed: usage of shadow fading model (yes or no). LTE_params.ChanMod_config.winner_settings.PathLossModel: path loss model function name (pathloss). LTE_params.ChanMod_config.winner_settings.PathLossOption: Available options are CR_light, CR_heavy, RR_light, RR_heavy. CR=Corridor-Room, RR=Room-Room NLOS. LTE_params.ChanMod_config.winner_settings.RandomSeed: sets a random seed. Can be left empty. LTE_params.ChanMod_config.winner_settings.UseManualPropCondition: whether to use manual propagation condition (LOS/NLOS) settings or not (yes or no). If not, the propagation condition is drawn from probabilities. LTE_params.ChanMod_config.corr_coefRX: correlation between the receiver antennas. Only compatible with block fading ltering. LTE_params.ChanMod_config.corr_coefTX: correlation between the transmitter antennas. Only compatible with block fading ltering. LTE_params.ChanMod_config.time_correlation: sets whether the channel realizations are time-correlated or not. correlated or independent. This parameter is also related with LTE_params.N_seed_reset.

E. Scheduler parameters Conguration of the scheduler parameters. LTE_params.scheduler.type: type of scheduler. Already numerous schedulers are supported (see [21] and [22]) the most important being: round robin: round robin scheduler without support of UE feedback. fixed: round robin scheduler but with support of UE feedback (MCS and MIMO preprocessing are adapted). best cqi: scheduling decisions are taken according to the reported channel quality, such as to maximize the system throughput. proportional fair: proportional fair scheduler. Other supported schedulers are (note: these are not thoroughly tested for all congurations): max throughput: max. throughput scheduler based on a linear program max min: max. min. scheduler var fair: throughput suboptimal scheduler with variable, adjustable fairness (it utilizes the parameter LTE_params.scheduler.fairness) variable fair: throughput optimal scheduler with variable, adjustable fairness alpha fair: scheduler with fairness indirectly set by the parameter LTE_params.scheduler.alpha All those schedulers assume full buffer simulations. The only scheduler that supports the trafc models is the RoundRobin_Traffic scheduler. LTE_params.scheduler.assignment: either static, semi static or dynamic. Whether the scheduler will statically assign or dynamically assign CQIs and other params. The best cqi scheduler is capable of assigning MCSs dynamically. The semi static scheduler adapts automatically the precoder and layer number according to the PMI and RI feedback (use this one in conjunction with closed loop spatial multiplexing and activated PMI and RI feedback). The semi static scheduler assigns resource blocks in a round robin fashion. LTE_params.scheduler.fixed_scheduler_assignment: this option is used only for the fixed scheduler. In that case, the parameter is a vector of length LTE_params.nUE containing the number of RBs allocated to each user. The total number or assigned RBs must be equal or less than the number of RB available in the specied bandwidth: eg. for 2 UEs and 1.4 MHz [4 2] would assign four RBs to the rst user and two to the second. LTE_params.scheduler.cqi: the CQI the scheduler is going to use when transmitting data. When set to set, when in the static case, the CQI value to use will be read from the cqi_i variable, which is to be set from the script le that launches the simulation (eg. LTE_sim_batch). This is only used with the round robin and fixed scheduler in static mode. LTE_params.scheduler.PMI: sets the Precoding Matrix Indicator, if no feedback is employed.

VI. R EFERENCE SIMULATION RESULTS INCLUDED Besides the one mentioned on Section II, the simulator is provided with some reference simulation results, which can be compared with performance curves from 3GPP RAN documents such as [23] to cross-check the results of the simulator. The example simulation results distributed with the simulator are described below. plot_reference_BLER_curves_r655 produces the plots in Figure 4, showing Block Error Ratio (BLER) and throughput for an Additive White Gaussian Noise (AWGN) simulation using the MCS dened in [11] (CQIs 1-15). plot_R1071967_throughput_curves_r553 shows the BLER and throughput curves obtained from performing AWGN simulations with the MCSs specied in R1-071967, page 16 [23], as seen in Figure 5. The MCS, including the CQI value used in the simulator, the modulation and Effective Code Rate (ECR) used in each of the simulations are shown in Table III.
CQI 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 Modulation 4QAM 4QAM 4QAM 4QAM 4QAM 4QAM 4QAM 4QAM 4QAM 4QAM 16QAM 16QAM 16QAM 16QAM 16QAM 16QAM 16QAM 16QAM 16QAM 64QAM 64QAM 64QAM 64QAM 64QAM 64QAM 64QAM 64QAM ECR [23] 1/9 1/6 0.21 1/4 1/3 0.42 1/2 0.58 2/3 0.73 0.43 0.46 1/2 0.54 0.58 0.61 2/3 0.73 4/5 0.58 0.62 2/3 0.70 0.74 4/5 0.85 0.90 ECRx1024 114 171 215 256 314 430 512 594 683 748 440 471 512 553 594 625 683 748 819 594 635 683 717 758 819 870 922

CQI 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Modulation 4QAM 4QAM 4QAM 4QAM 4QAM 4QAM 16QAM 16QAM 16QAM 64QAM 64QAM 64QAM 64QAM 64QAM 64QAM

ECR 0.0762 0.1172 0.1885 0.3008 0.4385 0.5879 0.3691 0.4785 0.6016 0.4551 0.5537 0.6504 0.7539 0.8525 0.9258

ECRx1024 [11] 78 120 193 308 449 602 378 490 616 466 567 666 772 873 948

TABLE III MCS S USED IN THE LTE_ S I M _ B A T C H _R1_07196 ( LEFT ) AND LTE_ S I M _ B A T C H _ Q U I C K _ T E S T ( RIGHT ) M ATLAB SCRIPTS

10

BLER, 1.4MHz, SISO AWGN, 5000 subframes

throughput, 1.4MHz, SISO AWGN, 5000 subframes

throughput [Mbps]

10

-1

10

-2

10

-3

CQI 01 CQI 02 CQI 03 CQI 04 CQI 05 CQI 06 CQI 07 CQI 08 CQI 09 CQI 10 CQI 11 CQI 12 CQI 13 CQI 14 CQI 15
-10 -5 0 5 10 SNR [dB] 15 20 25

5 CQI 01 CQI 02 CQI 03 CQI 04 CQI 05 CQI 06 CQI 07 CQI 08 CQI 09 CQI 10 CQI 11 CQI 12 CQI 13 CQI 14 CQI 15 -10 -5 0 5 10 SNR [dB] 15 20 25

BLER

-15

0 -15

(a) BLER, AWGN no HARQ Fig. 4. Reference BLER and throughput plots for the 15 MCSs dened in [11]

(b) Throughput, AWGN, no HARQ

10

BLER, 1.4MHz, SISO AWGN, 5000 subframes

throughput, 1.4MHz, SISO AWGN, 5000 subframes

5
QPSK, R=1/9 QPSK, R=1/6 QPSK, R=0.21 QPSK, R=1/4 QPSK, R=1/3 QPSK, R=0.42 QPSK, R=1/2 QPSK, R=0.58 QPSK, R=2/3 QPSK, R=0.73 16QAM, R=0.43 16QAM, R=0.46 16QAM, R=1/2 16QAM, R=0.54 16QAM, R=0.58 16QAM, R=0.61 16QAM, R=2/3 16QAM, R=0.73 16QAM, R=4/5 64QAM, R=0.58 64QAM, R=0.62 64QAM, R=2/3 64QAM, R=0.70 64QAM, R=0.74 64QAM, R=4/5 64QAM, R=0.85 64QAM, R=0.90 10 5 0 5 SNR [dB] 10 15 20 25

10

4
throughput [Mbps]

10

10

15

0 15

QPSK, R=1/9 QPSK, R=1/6 QPSK, R=0.21 QPSK, R=1/4 QPSK, R=1/3 QPSK, R=0.42 QPSK, R=1/2 QPSK, R=0.58 QPSK, R=2/3 QPSK, R=0.73 16QAM, R=0.43 16QAM, R=0.46 16QAM, R=1/2 16QAM, R=0.54 16QAM, R=0.58 16QAM, R=0.61 16QAM, R=2/3 16QAM, R=0.73 16QAM, R=4/5 64QAM, R=0.58 64QAM, R=0.62 64QAM, R=2/3 64QAM, R=0.70 64QAM, R=0.74 64QAM, R=4/5 64QAM, R=0.85 64QAM, R=0.90
10 5 0 5 SNR [dB] 10 15 20 25

BLER

(a) BLER, AWGN no HARQ Fig. 5. Reference BLER and throughput plots for the 27 MCSs dened in [23]

(b) Throughput, AWGN, no HARQ

VII. R EPRODUCING THE RESULTS / PLOTS PRESENTED ON A PAPER One of the main points of the simulator is to allow you to reproduce and review our results, as well as the algorithms that produce them. For each of the following publications a script is prepared in the folder paper scripts that will reproduce the results in the corresponding paper: Currently no papers presenting results obtained with the LTE-A simulator are published. The Vienna LTE Link Level Simulator contains several scripts for reproducing published results. VIII. E XTENSIONS OF THE V IENNA LTE-A L INK L EVEL S IMULATOR COMPARED TO THE V IENNA LTE L INK L EVEL S IMULATOR The Vienna LTE-A Link Level Simulator is a direct extension of the Vienna LTE Link Level simulator. It contains all the features and transmit modes implemented in the LTE simulator. The LTE simulator supports standard compliant transmission from up to four transmit antennas in one of the following transmit modes: Single antenna port scheme (mode 1): transmission from a single antenna port only. Transmit diversity scheme (mode 2): transmission from two antenna ports utilizing the Alamouti transmit diversity scheme. Open loop transmit diversity scheme/large delay CDD scheme (mode 3): transmission from 2 or 4 antenna ports, utilizing large delay Cyclic Delay Diversity (CDD). Closed loop spatial multiplexing (mode 4): transmission from 2 or 4 antenna ports, utilizing UE feedback for choosing the appropriate precoding matrix. Corresponds to the single and dual layer transmission schemes dened in the LTE spec. These four transmission modes all rely on the LTE reference signal structure, consisting of Cell Specic (CS) Reference Signals (RS). In addition to these transmission modes, the LTE-A simulator supports the following modes: MUMIMO based on the LTE codebook (mode 5): this mode allows to share spatial transmission layers between users. It is also based on CS-RS. Just a dummy mode is implemented here, assuming: A round robin scheduler (fixed MUMIMO) that equally distributes time, frequency and space resources to users. Only benchmark receivers are implemented, assuming full channel knowledge. Here each user decodes all spatial layers and throws away the ones that are not indended for him. Transmission modes based on UE specic RS and Channel State Information (CSI) RS Non-standard dened precoding (mode 8): ZF beamforming is implemented as an example, utilizing the ZF_MUMIMO scheduler. Up to eight layer transmisson (mode 9): this mode is the LTE-A counterpart of the LTE modes 4 and 5. It allows fast switching between SUMIMO and MUMIMO with up to eight transmit antennas. For MUMIMO the same dummy scheme as for mode 5 is implemented (employing the fixed MUMIMO scheduler), and the SUMIMO scheme is based on the UE feedback algorithms presented in [9]. More advanced features of LTE-A like e.g. carrier aggregation and basestation cooperation are not implemented in the rst release, but the structure of the simulator allows to include such features as well (by implementing appropriate schedulers,...).

10

IX. SNR DEFINITION In a MIMO-Orthogonal Frequency Division Multiplexing (OFDM) transmission system the SNR is the measurement for channel quality information and is a key factor of link error prediction. There are different measures or calculation procedures for the SNR in SISO and MIMO systems. For SISO systems the SNR can be viewed as receiving SNR, that is, received signal-to-noise ratio before the detector. Usually post-detection Signal to Interference and Noise Ratio (SINR) is used for MIMO link prediction. It is representing the SNR after combining in the receiver and measures the likelihood that a MCS will be decoded successfully. The LTE Link Level simulator SNR is dened as follows: T N 1 BS_output.y_tx contains the Tx-signal vector x = [x1 , . . . , xNT ] C T , where xk C with k [1, . . . , NT ] is the Tx-symbol sent from the k-th Tx-antenna (NT . . . number of Tx-antennas). We get 2 the total Tx-power, x = trace(Rx ) = trace(E xxH ) = 1 2 the Tx-power per Tx-antenna, xk = E{|xk |2 } = 1/NT 2 ChanMod_output.H denes the channel matrix H, with H F = NT NR (NR . . . number of Rx-antennas) BS_output.cell_genie.v denes the noise vector v with respect to the size of the Fast Fourier Transform (FFT) (LTE_params.Nfft) and the number of subcarriers (LTE_params.N_tot) before the detector, where vec(v) 2 CN (0, v I) 2 BS_output.cell_genie.n denes the noise vector n after the FFT, where vec(n) CN (0, n I) ChanMod_output.y_rx contains the Rx-signal vector y = Hx + v we get the receive SNR (before the detector), pret = where the SNR after the FFT becomes, postt = NR 1 Hx 2 F = = 2 2 2 NR n NR n n NR 1 Hx 2 F = = 2 2 2 NR v NR v v

2 2 The difference between pret and postt is given by the relation v /n = Nt /Ntot (using a FFT-size of 128 and 72 subcarrriers, we get Nt /Ntot = 128/72 = 1.778 2.5 dB). We recommend to use the SNR after FFT (postt ) when doing simulations over SNR, as this allows direct comparisons to theoretic results.

X. N OTE ON PARALLEL SIMULATIONS AND RANDOM NUMBER GENERATION Note that when using parallel simulations the random number generator will return the same sequence of numbers in each of the parallel-running M ATLAB labs. Since the parfor loop is over SNR values, this would mean that each SNR iteration is in principle identical to the other ones just with a different noise level (this may not be the case depending on the circumstances, though). If this is not adequate for your needs, you may need to modify the code regarding the RandStream initialization. Below is a code that illustrates what happens when using RandStreams in parallel mode: sim_length=15; n_sims = 10; output = zeros(n_sims,sim_length); for sim_=1:n_sims a_RandStream = RandStream(mt19937ar,Seed,0); matlabpool open parfor t_=1:sim_length pause(1); %simulate doing something output(sim_,t_) = rand(a_RandStream); end matlabpool close end output

11

XI. C HANGELOG Changelog of the Vienna LTE-A Link Level simulator: v.1.1r100, 2011-09-27 First release of the Vienna LTE-A Link Level simulator. The Vienna LTE-A Link Level simulator is an extension of the Vienna LTE Link Level simulator. It is built around Version 1.7 of the LTE simulator. Therefore the changelog of the Vienna LTE Link Level simulator until this version is also included: v.1.7r???, 2011-09-?? MEX les are now compiled with Microsoft Visual C++ 2010, so the appropriate runtime les are needed. Refer to Section XIII for more information on how to get them. Fixed typos in the LTE_sim_batch_BLER_curves le that reproduces the AWGN BLER curves. BICM capacity curves are now loaded by default not from a le, but from a hardcoded source. Loading from a le is still supported, though. Fixed error in the rate matching that caused an incorrect number of bits than otherwise expected by the layers for cases of bandwidth higher than 1.4 MHz and certain resource block allocations. Changed the layout of the load_parameters le. It was getting so long it was no longer readable. Moved the individual conguration parameters for each simulation type to separate object les in the simulation_config package. Fixed discrepancy initialization in the gold code generation MEX le. Thanks to Srikanth Isanaka for pointing out this bug and Mitsuo Sakamoto for generously providing an updated version of the gold code generation code. v.1.6r917, 2011-01-19 Fixed error in LTE_sim_results_plots when there is no frequency offsett estimation. Fixed signaling bug for TB sizes smaller than 40 bits. Thanks to Wang Dongming (Southeast University, China) for pointing out this bug. Fixed bug in the new implementation of the rate matching. The LLR repetition was not taken into account, therefore affecting performance. v.1.5r811, 2010-11-05 Added experimental multi-base station simulation support. Still in very early development stages. Reimplementation in M ATLAB of most of the rate-matching procedures. The code is optimized enough so that MEX functions are now much less needed. This should improve code readability and maintainability. Added Companies (no matter prot-oriented or not) are not allowed for free usage and have to contact the licensor before usage. to the license agreement (Section A). Fixed bug in the xed scheduler. The PBCHsyms variable was being used, thus causing the simulation to crash when the scheduler function was called. The CQI used for each Transport Block (TB) transmission in now stored in the results trace (used_CQI variable in ueSpecificTraces, in turn stored in simulation_results). Added trace to track the codeblock BLER. The ACK value is the decimal representation of the codeblock ACK vector for each codeword. v.1.4r715, 2010-07-12 Fixed bug in the channel coding process. According to [14]: If the code to be encoded is the 0-th code block and the number of ller bits is greater than zero, i.e. F > 0, then the encoder shall set ck = 0, k = 0, . . . , (F 1) at its input and shall set d0 =<NULL>, k = 0, . . . , (F 1) and d1 =<NULL>, k = 0, . . . , (F 1) at its output. These bits were k k erroneously set to 0 instead of <NULL>. In LTE_rx_DLSCH, the rst BS_signaling.TB_segmentation.F LLR values of d(0) and d(1) are replaced by -LTE_params.UE_config.LLR_clipping. Thanks to Wu Gaojin (University of Posts and Telecommunications, China) for pointing out this bug. Fixed bug in the RX rate matching process that caused a crash when no ller (<NULL>) bits were present (ie. UE_signaling.turbo_rate_matcher(stream_index).null_positions{i} was empty). Thanks to Narciso Garca Cano (Universidad Polit cnica de Madrid, Spain) for helping us out with this bug. e The Gold sequence generation was based on version 8.2.0 of the standard. Between the latest release and v8.2.0., pseudo random sequence generator was updated to include Nc [15]. Thanks to Mitsuo Sakamoto (Couei Corp, Japan) for helping us out with this bug and kindly providing an updated version of the Gold Sequence generation that is also 40x faster. Added comment on issues that may arise from using random number generators and parallel simulations. Added support of user equipment feedback (CQI, RI and PMI). The feedback values are evaluated as described in [6] (added a script to allow reproduction of the gures in this paper). CQI feedback is based on ESM. MIESM and EESM are supported for SINR averaging. The feedback calculation is just tested for the ZF receiver. The receiver for 4x1 and 4x2 TxD mode was modied. The performance is unchanged, but execution speed has been greatly improved.

12

v.1.3r620, 2010-02-18. Fixed bug in the turbo decoding process. The second decoder was initialized to the interleaved systematic bits instead of zero. Due to this there is a small performance improvement in the order of 0.2 dB. Thanks to Klaus Hueske (Technische Universit t Dortmund Information Processing Lab, Germany) for helping us out with this bug. a Fixed errors in the way the xed scheduler object was called that made it impossible to use it. Thanks to Tommaso Balercia (Comneon GmbH, Germany) for pointing out this bug. Changed structure of the simulations parameters loading le in order to decrease the number of conguration les. Added doubly dispersive channel estimation with scalable complexity [24] and related les needed to reproduce the gures from the paper. Added precoding matrix indicator feedback support as described in [25] and related les needed to reproduce the gures from the paper. v.1.2r553, 2009-12-25. Minor bugxes and improvements. Fixed bug in the rate matching process that caused the <NULL> bits inserted during the sub-block interleaving process to be treated as zeros (see [14]). Due to this x the performance of the channel coding is slightly improved. Reference and example simulations included with the simulator have been updated accordingly. Thanks to Ching Hsiang Wu (Institute for Information Industry Networks and Multimedia Institute, Taipei) for pointing out this bug. Fixed bug that caused poor performance of the ZF receiver. This was due to a bad noise scaling being passed on to the demapper. After the x the ZF and Soft Sphere Decoder (SSD) receivers have the same performance for the SISO case. Since according to the MathWorks,the seqgen function will be removed from future versions of the Communications Tollbox software, calls to seqgen.pn have been substituted by calls to commsrc.pn. See Section XIII for issues this change may cause. Fixed bug that caused performance decrease for fast fading simulations at high user velocities. For the decoding of the last seven OFDM symbols, the channel of the rst seven OFDM symbols was used. Added support of the Winner II + channel model [26]. See Section XIV on more information on how to enable this functionality. v.1.1r450, 2009-08-25. Minor bugxes and improvements. Added Parallel and Distributed Toolbox support (parfor). Added best CQI (only for SISO and xed schedulers. v.1.0r400, 2009-05-15. First publicly available version of the LTE Link Level Simulator. XII. R EFERENCING

A version of the LTE Link Level Simulator paper is available in our publication data-base here. If you are using the simulator for your scientic work, please use the refence below:
@ARTICLE{JASP2011, author = {C.Mehlf\"uhrer and J.Colom Ikuno and M.Simko and S.Schwarz and M.Wrulich and M.Rupp}, title = {The Vienna {LTE} Simulators - Enabling Reproducibility in Wireless Communications Research}, journal = {{EURASIP} Journal on Advances in Signal Processing}, year = {2011}, volume = {Vol. 2011}, pages = {1-13} } C. Mehlf\"hrer, J. Colom Ikuno, M. Simko, S. Schwarz, M. Wrulich, M. Rupp, The Vienna LTE Simulators - Enabling Reproducibility in Wireless Communications Research, EURASIP Journal on Advances in Signal Processing, Vol. 2011, pages 1 - 13, 2011.

XIII. K NOWN ISSUES

The LTE simulators make use of the new Object-Oriented capabilities of Matlab (available since R2008a), the simulators will not run under older Matlab releases without extensive changes. Please note that MEX-les generated using Microsoft Visual C++ 2010 require that Microsoft Visual Studio 2008 run-time libraries be available on the computer they are run on. The runtime les can be downloaded here (x86) or here (x64). In order to be able to use the parallel version of the simulator (when setting LTE_params.simulation_type to parallel, you need the parallel toolbox (included by default with M ATLAB r2009a and above or as an add-on with previous versions). It will not work if you dont have the toolbox, just crashing the moment the matlabpool function is called.

13

In M ATLAB versions prior to r2009a the code may not work, as the commsrc.pn function does not exist. You will need to replace every call to commsrc.pn with a call to seqgen.pn in order to run the simulator. No change in the arguments is needed. Such changes should be applied to the code in the LTE_common_gen_Synchronization_Signal and LTE_common_gen_Reference_Signal functions. In the LTE_rx_turbo_decode function, only the max-log-map decoder type has been tested. The decoder_type variable is used as input an congures the SISO decoder function, which is part of [27]. It was pointed out that in [12], the phase is not different for each sinusoid. We are using a modied version [13]. XIV. U SING THE W INNER P HASE II CHANNEL MODEL REFERENCE IMPLEMENTATION

Starting with v.1.2r553, it is possible to use channels generated with the publichly-available M ATLAB implementation of the WINNER Phase II Channel Model [26]. Since the code is distributed under the GNU GPL, its les are not included in the simulator release. In order to use to be able to use it, you will have to download it yourself. For this, go to the WINNER Phase II Model website, download the WIM2_3D_ant_ver064_220908.zip le and unzip the .mat les in the ./Winner Channel Model folder. XV. Q UESTIONS For questions please check our forum, where you will be able to post your questions/comments/bug reports. It makes it easer for you to see what other people asked and also makes it easier for us to answer you (when we have time). XVI. M AILING L IST If you want to receive information about future updates you can subscribe to our LTE simulator mailing list that you can change the display language to english in the selection panel to the right. XVII. T HE P EOPLE ( SO FAR ) BEHIND THE DEVELOPMENT OF THE SIMULATOR

here. Note

Dagmar Bosanska Josep Colom Ikuno Govinda Lilley Michal Simko Christian Mehlf rer u Michael Meidlinger Martin M ller u Stefan Pendl J rg Reitterer o Markus Rupp Stefan Schwarz Qi Wang Martin Wrulich XVIII. L ICENSE AGREEMENT

These terms (license for the LTE link level simulator) refer to the use of the LTE-A link-level simulator (the Original Work), developed by the Institute of Telecommunications, Vienna University of Technology (the licensor). A. Academic Usage Academic Usage in the context of this license describes the use of the Original Work in scientic projects without any reimbursement or nancial claims that bear on results derived by the Original Work, but subject however to the restrictions provided for in Clauses B and H hereinbelow. The main goal in the sense of Academic Usage shall be to obtain cientically signicant results that can be used for publication. Academic Usage of the Original work will only be granted after this license agreement is: 1) printed on letterhead of university department or research institute, 2) signed by the department head, including date and stamp of the department, 3) then scanned and, including the above, and emailed to: Univ.Prof. Markus Rupp (mrupp@nt.tuwien.ac.at). Companies (no matter prot-oriented or not) are not allowed for free usage and have to contact the licensor before usage.

14

B. Grant of copyright license Licensor grants You a worldwide, royalty-free, non-exclusive, non-sublicensable license, restricted to non-commercial use, for the duration of the copyright, to install the Original ork and any Derivative Works thereof on one personal computer. The license allows You to: 1) Use the Original Work only for Academic Usage. Any usage of the Original Work, entirely or in part or modied, requires the proper citation, e.g. as reference in a publication. 2) Translate, adapt, alter, transform, modify, or arrange the Original Work, thereby creating derivative works (Derivative Works) based upon the Original Work. Distribution, either royalty-free or commercially, in parts or in modied form of the Original Work, i.e. also of Derivative Works, is prohibited and not covered by Academic Usage. 3) Display results derived from the Original Work, or in modied form, publicly, without commercial usage. C. Grant of source code license The term Source Code means the preferred form of the Original Work for making modications to it and all available documentation describing how to modify the Original Work. Licensor agrees to provide a machine-readable copy of the Source Code of the Original Work along with each copy of the Original Work that Licensor distributes. Licensor reserves the right to satisfy this obligation by placing a machine-readable copy of the Source Code in an information repository reasonably calculated to permit inexpensive and convenient access by You for as long as Licensor continues to distribute the Original Work. D. Exclusions from license grant Neither the names of Licensor, nor the names of any contributors to the Original Work, nor any of their trademarks or service marks, may be used without express prior permission of the Licensor, except as expressly provided otherwise in Clause B1 hereinabove. Except as expressly stated herein, nothing in this License grants any license to Licensors trademarks, copyrights, patents, trade secrets or any other intellectual property. No license is granted to the trademarks of Licensor even if such marks are included in the Original Work. Nothing in this License shall be interpreted to prohibit Licensor from licensing under terms different from this License any Original Work that Licensor otherwise would have a right to license. E. Warranty of provenance and disclaimer of warranty Licensor warrants that the copyright in and to the Original Work is owned by the Licensor or is sublicensed to You under the terms of this License with the permission of the contributor(s) of those copyrights and patent rights. Except as expressly stated in the immediately preceding sentence, the Original Work is provided under this License on an AS IS BASIS and WITHOUT WARRANTY, either express or implied, including, without limitation, the warranties of noniinfringement, merchantability or tness for a particular purpose. THE ENTIRE RISK AS TO THE QUALITY OF THE ORIGINAL WORK IS WITH YOU. This DISCLAIMER OF WARRANTY constitutes an essential part of this License. No license to he Original Work is granted by this License except under this disclaimer. F. Limitation of liability Under no circumstances and under no legal theory, whether in tort (including negligence), contract, or otherwise, shall the Licensor be liable to anyone for any indirect, special, incidental, or consequential damages of any character arising as a result of this License or the use of the Original Work including, without limitation, damages for loss of goodwill, work stoppage, computer failure or malfunction, or any and all other commercial damages or losses. This limitation of liability shall not apply to the extent applicable law prohibits such limitation. G. Termination If, at any time, You infringe upon the grants of this License, it shall terminate immediately and You may no longer exercise any of the rights granted to You by this License. H. Open source and code under other license terms The original work also contains work licensed under other licenses other than the license for the LTE link-level simulator. The terms and conditions described in this document are only applicable to the parts of the Original Work not under other licenses. A list of the parts of the Original Work not under the license for the LTE link-level simulator can be found in Appendix I.

15

I. Appendix I The following parts of the original work are not under the terms of the license for the LTE link-level simulator, and are thus excluded from the terms and conditions stated by this license. The usage and adaptation of these sections for use with the original work is done in compliance with the license terms they are released under. Any translation, adaptation, alteration, transformation, modication, or further use of the hereinbelow stated parts of the original work must be done under the terms of the applicable licenses for that specic part, which are also included in the package. CRC calculations C-code: MEXed version of the code generated by pycrc [28]. Under the MIT License. Turbo and convolutional encoder/decoder C-code, from the Coded Modulation Library (CML) by Iterative Solutions [27]. Licensed under the GNU lesser General Public License.

Use

CRC generation and checking

Channel coding

c-source crc16.c crc16.h crc24a.c crc24a.h crc24b.c crc24b.h ConvEncode.c convolutional.h maxstar.h siso.h SisoDecode.c

M ATLAB functions LTE_common_crc16 LTE_common_crc24a LTE_common_crc24b

License

MIT License [4]

LTE_tx_convolutional_encoder LTE_rx_siso_decode

GNU Lesser GPL [3]

ACKNOWLEDGMENT The authors would like to thank the whole LTE research group for continuous support and lively discussions. This work has been funded by mobilkom austria AG, the Christian Doppler Laboratory for Wireless Technologies for Sustainable Mobility, as well as the Institute of Telecommunications Vienna University of Technology. The views expressed in this paper are those of the authors and do not necessarily reect the views within mobilkom austria AG. R EFERENCES
[1] [Online]. Available: http://www.nt.tuwien.ac.at/ltesimulator/ [2] C. Mehlf hrer, M. Wrulich, J. C. Ikuno, D. Bosanska, and M. Rupp, Simulating the long term evolution physical layer, in Proc. of the 17th European u Signal Processing Conference (EUSIPCO 2009), Glasgow, Scotland, Aug. 2009. [3] I. Free Software Foundation, GNU lesser general public license, version 2.1. [Online]. Available: http://www.gnu.org/licenses/lgpl-2.1.html [4] M. I. of Technology, MIT license. [Online]. Available: http://www.opensource.org/licenses/mit-license.php [5] Technical Specication Group Radio Access Network, Evolved universal terrestrial radio access (E-UTRA); LTE physical layer general description, 3rd Generation Partnership Project (3GPP), Tech. Rep. TS 36.201 Version 8.3.0, Mar. 2009. [6] S. Schwarz, C. Mehlfhrer, and M. Rupp, Calculation of the Spatial Preprocessing and Link Adaption Feedback for 3GPP UMTS/LTE, in Proc. IEEE Wireless Advanced 2010, London, UK, June 2010. [7] C. Mehlfuhrer, J. Colom Ikuno, M. Simko, S. Schwarz, M. Wrulich, and M. Rupp, The vienna lte simulators - enabling reproducibility in wireless communications research, EURASIP Journal on Advances in Signal Processing, vol. 2011, no. 1, p. 29, 2011. [Online]. Available: http://asp.eurasipjournals.com/content/2011/1/29 [8] 3GPP, Technical Specication Group Radio Access Network; Evolved Universal Terrestrial Radio Access (E-UTRA); Physical Channels and Modulation (Release 10), December 2010, [Online]. Available: http://www.3gpp.org/ftp/Specs/html-info/36211.htm. [9] S. Schwarz and M. Rupp, Throughput maximizing feedback for MIMO OFDM based wireless communication systems, in Signal Processing Advances in Wireless Communications SPAWC 2011, San Francisco, CA, June 2011. [10] G. Dietl, O. Labreche, and W. Utschick, Channel vector quantization for multiuser MIMO systems aiming at maximum sum rate, in IEEE Global Telecommunications Conference 2009, Dec. 2009, pp. 1 5. [11] Technical Specication Group Radio Access Network, Evolved universal terrestrial radio access (E-UTRA); physical layer procedures, 3rd Generation Partnership Project (3GPP), Tech. Rep. TS 36.213, Mar. 2009. [12] Y. R. Zheng and C. Xiao, Simulation models with correct statistical properties for rayleigh fading channels, Communications, IEEE Transactions on, June 2003. [13] T. Zemen and C. Mecklenbr uker, Time-Variant Channel Estimation Using Discrete Prolate Spheroidal Sequences, IEEE Transactions on Signal a Processing, vol. 53, no. 9, pp. 35973607, Sept. 2005. [14] Technical Specication Group Radio Access Network, Evolved universal terrestrial radio access (E-UTRA); multiplexing and channel coding, 3rd Generation Partnership Project (3GPP), Tech. Rep. TS 36.212, Mar. 2009. [15] , Evolved universal terrestrial radio access (E-UTRA); physical channels and modulation, 3rd Generation Partnership Project (3GPP), Tech. Rep. TS 36.211 Version 8.7.0, May 2009. [16] , Evolved universal terrestrial radio access (E-UTRA); user equipment (UE) radio access capabilities, 3rd Generation Partnership Project (3GPP), Tech. Rep. TS 36.306, 2009. [17] 3GPP, Technical Specication Group Radio Access Network; Evolved Universal Terrestrial Radio Access (E-UTRA); Physical layer procedures (Release 10), December 2010, [Online]. Available: http://www.3gpp.org/ftp/Specs/html-info/36213.htm. [18] ITU-R, Guidelines for evaluation of radio transmission technologies for IMT-2000, ITU-R, Tech. Rep. M.1225, 1997.

16

[19] Technical Specication Group GSM/EDGE Radio Access Network, Radio transmission and reception, annex c.3 propagation models, 3rd Generation Partnership Project (3GPP), Tech. Rep. TS 05.05 V.8.20.0 (Release 1999), 2009. [20] L. Hentil , P. Ky sti, M. K ske, M. Narandzic, and M. Alatossava. (2007) MATLAB implementation of the WINNER Phase II channel model ver1.1. a o a [Online]. Available: http://www.ist-winner.org/phase 2 model.html [21] S. Schwarz, C. Mehlf hrer, and M. Rupp, Low complexity approximate maximum throughput scheduling for LTE, in 44th Annual Asilomar Conference u on Signals, Systems, and Computers, Pacic Grove, California, Nov. 2010. [22] , Throughput maximizing multiuser scheduling with adjustable fairness, in International Conference on Communications ICC 2011, Kyoto, Japan, June 2011. [23] Alcatel-Lucent, DL E-UTRA performance checkpoint, 3GPP TSG-RAN1, Tech. Rep. R1-071967, 2007. [24] M. Simko, C. Mehlf hrer, M. Wrulich, and M. Rupp, Doubly Dispersive Channel Estimation with Scalable Complexity, in Proc. WSA 2010, Bremen, u Germany, Feb. 2010, accepted. [25] S. Schwarz, M. Wrulich, and M. Rupp, Mutual information based calculation of the precoding matrix indicator for 3GPP UMTS/LTE, in Proc. WSA 2010, Bremen, Germany, Feb. 2010, accepted. [26] L. Hentil , P. Ky sti, M. K ske, M. Narandzic, and M. Alatossava, MATLAB implementation of the WINNER Phase II Channel Model ver1.1, a o a December 2007. [Online]. Available: https://www.ist-winner.org/phase 2 model.html [27] I. Solutions, Iterative Solutions Coded Modulation Library (ISCML). [Online]. Available: http://www.iterativesolutions.com/ [28] T. Pircher, pycrc CRC calculator and C source code generator. [Online]. Available: http://www.tty1.net/pycrc/